Downloads · 30 days
0
genie-pong
sangramrout/genie-pong
genie-pong is a reinforcement learning model from sangramrout. Use it for the reinforcement learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
This repository contains trained checkpoints for the Genie world model, trained on the Pong game.
Access
Public
Updated Jan 3, 2026
Repo size
7.8 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt7.8 GB · 100%
At a glance
- Task
- Reinforcement Learning
- License
- mit
- Access
- Public
- Created
- Jan 3, 2026
- Updated
- Jan 3, 2026
- SHA
- 18e5872c
Model card
From the Hugging Face model README
Genie World Model - Pong
This repository contains trained checkpoints for the Genie world model, trained on the Pong game.
Models
| Model | Description | Size |
|---|---|---|
| Video Tokenizer | ST-ViViT encoder/decoder with 512-code VQ codebook | ~454 MB |
| Latent Action Model (LAM) | Learns 3-action discrete space (up/down/nothing) | ~6.2 GB |
| Dynamics Model | MaskGIT-style next-frame predictor | ~647 MB |
Usage
import torch
from huggingface_hub import hf_hub_download
# Download tokenizer
tokenizer_path = hf_hub_download(
repo_id="sangramrout/genie-pong",
filename="checkpoints/tokenizer/checkpoint_step_2288.pt"
)
# Load checkpoint
checkpoint = torch.load(tokenizer_path, map_location="cpu")
Training
See the GitHub repository for training code and details.
Architecture
- Video Tokenizer: 6-layer encoder, 8-layer decoder, 384 d_model
- LAM: 20-layer encoder/decoder (paper hyperparameters), 1024 d_model, 3-action codebook
- Dynamics: 8-layer transformer, 640 d_model, MaskGIT training
- Task
- Reinforcement Learning
- License
- mit
- Created
- Jan 3, 2026
- Updated
- Jan 3, 2026